Supervised Brain Network Learning Based on Deep Recurrent Neural Networks

Shijie Zhao,Li Xie,Lei Guo,Linwei Huang,Yaowu Chen,Tianming Liu,Junwei Han,Jinglei Lv,Yan Cui,Shu Zhang

doi:10.1109/access.2020.2984948

Abstract

Task-based functional magnetic resonance imaging (tfMRI) is a widely used neuroimaging technique in exploring brain networks and functions associated with cognitive behaviors. Traditionally, the general linear model (GLM) is the most popular method in tfMRI data analysis due to its simpleness and robustness. This model-driven method adopts a canonical hemodynamic response function (HRF) and its various derivatives to construct regressors in the design matrix and estimate changes in the tfMRI data. However, a possible limitation of current model-driven methods is that the HRF is fixed and non-adaptive which may overlook other diverse and concurrent brain networks. In order to overcome these limitations, we proposed a novel hybrid framework, supervised brain network learning based on deep recurrent neural networks (SUDRNN), to reconstruct the diverse and concurrent functional brain networks. Specifically, this hybrid framework first takes advantage of the great capacity of deep recurrent neural networks (DRNN) in modeling sequential data to learn the diverse regressors from real tfMRI data. After that, it utilizes the effective supervised dictionary learning (SDL) method to reconstruct both the task-related functional brain networks and other latent brain networks simultaneously. Extensive experiment results on different tfMRI datasets from Human connectome project (HCP) demonstrated the superiority of the proposed framework.

Highlights

Functional magnetic resonance imaging is one of the most popular noninvasive neuroimaging techniques in the study of neuroscience, experimental psychology, and brain disorders [1]–[3]
In [26], [34], supervised dictionary learning methods were proposed to detect task evoked brain networks and spontaneous brain networks simultaneously. These regressors are still borrowed from general linear model (GLM) method or generated solely from theory assumptions lacking flexibility and adaptability [26], [34]. Motivated by these impressive works in recurrent neural networks and supervised dictionary learning studies, we proposed a novel hybrid framework, supervised brain network learning based on deep recurrent neural networks (SUDRNN), to reconstruct the diverse and concurrent functional brain networks from task-based Functional magnetic resonance imaging (fMRI) (tfMRI) data
SUPERVISED BRAIN NETWORK LEARNING BASED ON DRNN After learned the meaningful and adaptive regressors from real tfMRI data, we employed the supervised dictionary learning method [26] to identify the diverse and concurrent functional brain networks in tfMRI data, forming a supervised brain network learning based on deep recurrent neural networks (SUDRNN) framework

Summary

Introduction

Functional magnetic resonance imaging (fMRI) is one of the most popular noninvasive neuroimaging techniques in the study of neuroscience, experimental psychology, and brain disorders [1]–[3]. There have been consistent effort in developing brain network reconstruction and modeling techniques including the general linear model (GLM)[9], [10], principal component analysis (PCA) [11], independent component analysis(ICA) [12], [13], sparse representation based methods(SR) [14]–[18]. Among all of these methods, GLM [9], [10] is the most popular methodology in tfMRI data analysis. The basic idea of this model-driven method is utilizing the hemodynamic response function (HRF) [19]–[21] and the experimental paradigm to construct

Methods

Results

Conclusion